The Myth of the SaaSpocalypse

In the Fall of 1999, I was sitting in a conference room at the Software & Information Industry Association, trying to write the industry's first formal definition of something the technology sector was practicing but hadn't yet named. A small group of us drafted a definition of “Software As A Service” to be included in a Trends Shaping the Digital Economy Report. It was among the first formal uses of what the world would come to know as SaaS. We thought we were defining a delivery model. We had no idea we were naming an era.

Twenty-five years later, pundits are writing SaaS's obituary under the banner of the SaaSpocalypse. The logic goes: AI agents will do the work that humans once did, companies will need fewer seats, and the subscription revenue model that powered a generation of enterprise software giants will quietly collapse. From where I sit running a venture firm in Chicago that connects large corporations and startups, watching this play out inside real companies, the numbers tell a different story.

I've watched this cycle before. Every major platform shift arrives with confident predictions that the incumbents are finished. IBM's mainframe software business was supposed to disappear when the PC arrived — it still generates billions in free cash flow. SAP was supposed to be dismantled by client-server, then the web, then mobile, then cloud; its revenues have grown more than sixtyfold since 1992. Platform transitions are real. Mass incumbent extinction is a much rarer outcome.

The Salesforce numbers from Q1 FY27 are worth examining. Revenue hit $11.1 billion, up 13% year over year, with subscription growth climbing from 9% to 12% in constant currency — a genuine reacceleration at more than $45 billion in annualized run-rate revenue. More telling is the composition: more than half of bookings for Agentforce and the Data 360 platform came from existing customers expanding their software spend alongside their existing licenses. 

The predicted mass defection hasn't materialized. That shouldn't surprise anyone watching from the middle of the country — the manufacturers, health systems, and logistics companies that have run on enterprise software for decades aren't looking for the exits. They're asking how to get more out of what they already have.

We see it in our own ecosystem. The corporations we partner with aren't rethinking their enterprise software stacks. They're asking us to find the startups building on top of them.

The doom narrative also misunderstands how AI reshapes software's competitive position. An agent embedded in a customer's workflow accumulates that customer's exceptions, quirks, and institutional logic over time. My TechNexus co-founder Terry Howerton calls this living proprietary data — information generated by product usage that's impossible to replicate from the outside. Companies with long customer histories and deeply embedded workflows find the AI transition accelerates their advantages rather than exposing them.

Midwest investors have built track records around exactly this thesis for years. Terry's framework for what makes an AI company fundable — proprietary data, deep domain expertise, and distribution embedded enough to become structural — maps closely onto what this region's best software businesses have always been built on. The current moment is proving that approach right.

Enterprise software is absorbing AI rather than being displaced by it. The application layer grows more valuable as agents require trusted, deeply embedded systems to function — infrastructure accumulated across years of customer relationships, compliance work, and institutional workflow knowledge that any new entrant would spend years replicating.

At TechNexus, we've spent nearly two decades connecting established corporations — many of them Midwest anchors — with the startups building the future. The companies navigating this AI moment most effectively are running both sides of that bridge, leaning on the data depth and workflow trust of their incumbent software stack while working with startups to build new AI capability on top. 

I helped define SaaS twenty-five years ago because subscription-delivered software represented something genuinely new about how software could be valued and accessed. The companies treating this AI moment as the next chapter of that model are already pulling ahead.

‍

Fred Hoch is co-founder and General Partner of TechNexus Venture Collaborative, a Chicago-based VC firm that helps leading corporations invest in ambitious startups. In 1999, he co-authored "Trends Shaping the Digital Economy" while serving as Director of the eBusiness Division at the Software & Information Industry Association (SIIA) — one of the first publications to formally define the SaaS model.